Noether normalisation of affinoid algebras #
Every nonzero affinoid algebra over a complete ultrametric field admits a finite injective algebra map from a Tate algebra. No reducedness, characteristic, or nontrivial valuation assumption is needed.
The argument follows Bosch–Güntzer–Remmert, Non-Archimedean Analysis, §6.1.2. Its
distinguished-variable induction adapts the central construction in
TauCeti.MvPowerSeries.finite_quotient_of_isMaximal: a nonzero relation becomes distinguished
after a change of variables, so Weierstrass division makes the algebra finite over the image of
a Tate algebra with one fewer variable. Applying induction to that image retains nilpotents.
The changes of variables are isometries of Gauss norms, so for a presentation
π : K⟨X₁, …, Xₙ⟩ → A the normalisation can be chosen of the form π ∘ ι for a Gauss-norm
preserving map ι : K⟨Y₁, …, Y_d⟩ → K⟨X₁, …, Xₙ⟩ (AlgHom.exists_norm_eq_finite_injective_comp).
In particular it sends the Gauss unit ball into the image under π of the Gauss unit ball
(AlgHom.exists_finite_injective_map_unitClosedBall_le). This compatibility relates the
supremum seminorm of A to integrality over the presentation ring π(K⟨X₁, …, Xₙ⟩°), and lets
coefficients over K⟨Y₁, …, Y_d⟩ be lifted along π without increasing their Gauss norms.
Affinoid Noether normalisation. Every nonzero affinoid algebra over a complete ultrametric field is finite over an injectively embedded finite-variable Tate algebra. Finiteness uses the algebra structure induced by the displayed map.
Affinoid Noether normalisation through a presentation. For a Tate presentation π of a
nonzero algebra there is a map ι : K⟨Y₁, …, Y_d⟩ → K⟨X₁, …, Xₙ⟩ of Tate algebras, preserving
Gauss norms, such that π ∘ ι is finite and injective.
Affinoid Noether normalisation adapted to a presentation. For a Tate presentation π of
a nonzero algebra, a finite injective map g from a Tate algebra can be chosen to send the
Gauss unit ball into the image under π of the Gauss unit ball.